Humboldt-Universität zu Berlin - CALLIDUS

Outcomes of the Workshop

One thing has now become quite clear: We don't have much to build on. Neither can we fall back on a high-quality annotated and scholastically interesting Latin corpus, nor do the existing POS taggers give hope for a reliable (automatic) solution to the problem.

Decisions Regarding the Prototype of the Machina Callida

Since LangBank and Laudatio are no longer eligible for the project for various reasons, Perseus or, if applicable, PROIEL will be used for the annotated text corpus. Accordingly, only a small corpus (four letters, two each from Cicero and Pliny) can form the basis, since they have to be annotated by hand. Automatic POS tagging is put back as a problem.

The main focus of the work is the creation of a working prototype, with which experiences can be gathered in school starting April 2018.


Open Questions

Questions remain open for the programming: Which framework should be used? Is the import into Moodle to be implemented as push or pull? Will the application be designed web-based and/or app-based (or both)? Will the software be developed from the end (user) or the beginning (corpus)? ...



❮ back       forward❯